Skip to content

Step 5: Modify your data

If necessary, you can update a dataset after publication to correct errors, improve documentation, add files, or incorporate additional information.

DataverseNO uses versioning to ensure that changes can be tracked while preserving access to earlier versions of the dataset.

This step explains:

Why modify a dataset?

You may wish to create a new version of a dataset in order to:

  • Correct metadata.

  • Update the README file.

  • Add new data files.

  • Replace files with improved versions.

  • Add information about related publications.

  • Address recommendations from users or collaborators.

Updating a dataset allows you to improve its quality and usefulness while preserving the scholarly record.

How versioning works

To make changes to a published dataset, log in to DataverseNO, navigate to the dataset you want to update, and click Edit Dataset on the dataset landing page. To upload a new version of a file, you should first delete the old one.

When changes are made:

  • A new draft version of the dataset is created.

  • The draft must be submitted for review.

  • A curator reviews the changes.

  • The new version is published.

To support transparency, reproducibility, and an efficient review process, we recommend clearly documenting all changes from the previous version in the README file. This information should preferably be recorded in the dedicated Version history section of the README file template, under the question “Is this an updated version of a dataset published on DataverseNO?”.

What happens after you submit a new version?

New versions follow the same review process as new datasets.

As explained in Step 3: Curation and publishing, curators review submitted datasets to help ensure that they remain well documented, understandable, and reusable. When reviewing a new version of a dataset, particular attention is paid to the changes introduced since the previous version.

Earlier versions remain available after publication of the new version. This ensures that users can always access the exact version of the data that was used in earlier research.

What happens to the DOI and citation?

DataverseNO uses Version Control to track all changes made to a published dataset. This ensures transparency and allows users to identify exactly which version of the dataset was used in a particular study.

The dataset DOI remains the same when a new version is published. This provides a stable identifier for the dataset throughout its lifecycle.

Changes are assigned version numbers. Depending on the nature of the changes, a version may be released as either a major version or a minor version.

Major versions

Major versions are typically used when data files are added, removed, replaced, or otherwise modified.

Examples:

V1 -> V2
V2 -> V3

Major version changes are reflected in the dataset citation.

Example:

Hansen, L. M., Berg, S. E., & Nilsen, T. R. (2026). Bird observations from Northern Norway, 2018-2024 (Version 1) [Data set]. DataverseNO. https://doi.org/10.18710/ABCDE1

becomes:

Hansen, L. M., Berg, S. E., & Nilsen, T. R. (2027). Bird observations from Northern Norway, 2018-2024 (Version 2) [Data set]. DataverseNO. https://doi.org/10.18710/ABCDE1

Minor versions

Minor versions are typically used for smaller updates that do not affect the underlying data files.

Examples include correcting or expanding metadata, for example adding information to the Related Publication field.

Examples:

V1 -> V1.1
V1.1 -> V1.2

Minor version changes are tracked by the repository but are normally not reflected in the dataset citation. In these cases, the citation remains at the major version level.

For example, even if the internal repository version changes from V1 to V1.1, the recommended dataset citation will continue to refer to V1.

Can a published dataset be deleted?

Published datasets cannot normally be deleted.

This is because published datasets have persistent identifiers, such as DOIs, and form part of the scholarly record.

Deaccessioning

In exceptional cases, access to files in a published dataset may be removed. This process is called deaccessioning. Deaccessioning may be considered only when there is a compelling reason, for example if the dataset does not meet DataverseNO deposit criteria, contains malware, violates copyright, contractual obligations, legal requirements, research ethics, or involves research misconduct.

When a dataset is deaccessioned:

  • The data files are no longer publicly accessible.

  • The dataset metadata remains visible.

  • The README file remains accessible.

  • The dataset DOI remains part of the scholarly record.

If you believe that a published dataset should be deaccessioned, please contact your local user support.

Need help?

If you are unsure whether a dataset should be updated, how versioning will affect citation, or whether deaccessioning may be appropriate, please contact your local user support.

You've reached the end of the deposit workflow

Congratulations! You now know how to:

  • Prepare your data.

  • Deposit your data.

  • Get your data published.

  • Refer to your data.

  • Update your dataset when needed.

For additional guidance, see: